- Startup Spells 🪄
- Posts
- Oreo's $4B Marketing Playbook: Always Be Launching
Oreo's $4B Marketing Playbook: Always Be Launching
PLUS: Subtle Tricks Shopping Sites Use to Make You Spend More
Oreo's $4B Marketing Playbook: Always Be Launching
Oreo has mastered the art of perpetual relevance through a relentless cadence of limited-edition flavors and high-profile collaborations that keep the brand constantly in the spotlight.
Their strategy demonstrates how product variations can transform a century-old cookie into a marketing powerhouse.
The brand introduces new flavors every few months, creating regular touchpoints with consumers and maintaining cultural relevance.
The most effective marketing strategy is launching new products every few weeks or months. Running this cycle of launch every 4 to 6 weeks is how you get maximum hits at the bat.
A big company like Oreo goes from initial ideation to market release anywhere from 6 months to 3 years.
Flavor Innovation as Marketing Moments
Each new Oreo variant becomes a newsworthy event that drives conversation and consumer excitement:
Birthday Cake Oreo (2012): Released for Oreo's 100th anniversary with cake-flavored creme and colorful sprinkles; popularity led to permanent status
Candy Corn Oreo (2012-2016): Target-exclusive seasonal offering with orange/yellow creme that generated Halloween buzz
Fruit Punch Oreo (2014): Bold experimental flavor that sparked online debate and media coverage
January 2025 Multi-Launch: Six simultaneous new products including Game Day Oreos, Irish Creme Thins, and Oreo Loaded

Oreo Multiple Flavors
Strategic Brand Collaborations That Amplify Reach
Partnerships with entertainment properties and celebrities allow Oreo to tap into established fan bases and create cross-promotional opportunities.
Auntie Anne’s Oreo Topped Nuggets (April 2025): Salty pretzel nuggets drizzled with cookies-and-creme icing and topped with crushed Oreo pieces, launched nationwide after a successful test run.
Post Malone Oreo (Feb 2025): Featured innovative "twisted creme" combining salted caramel and shortbread flavors
Pokémon Oreo (2021): Created collector culture with 16 different Pokémon-stamped cookies
Game of Thrones Oreo (2019): House sigil designs timed with final season premiere
Supreme Oreo (2020): Red wafers featuring the streetwear brand's logo became instant collector's items

Oreo x Auntie Anne's Collab
Descript’s Repeated Product Hunt Launches Drive Growth and Engagement
Like Oreo, Descript leverages frequent launches to stay top-of-mind and reach new audiences. Their cadence of Product Hunt releases turns each new feature into a marketing moment and keeps users engaged.
Descript’s launch cadence fuels brand momentum. Multiple Product Hunt launches act as growth engine:
Each launch introduces Descript to new users and re-engages existing ones.
Frequent updates showcase ongoing innovation, building trust and excitement.
Product Hunt recognition (multiple #1 and top-5 rankings) amplifies reach and credibility.
As with Oreo, this steady drumbeat of launches keeps Descript relevant and top-of-mind in a crowded market.

Descript's Multiple Product Hunt Launches
The Business Genius Behind the Flavor Frenzy
This continuous launch strategy delivers multiple benefits that extend far beyond the sales of individual limited editions:
Media Magnetism: Each new flavor generates free press coverage and social media conversation
Core Product Growth: Limited editions drive foot traffic to the cookie aisle, where consumers often purchase the original Oreo alongside special flavors
Real-World R&D: These launches essentially fund future product development. Once the products are validated, they can be rolled into the company's long-term product roadmap.

Oreo Multiple Packages
Oreo's strategy proves that in today's market, the simple act of launching something new-consistently and creatively-can be the most powerful marketing tool a brand possesses.
Top Tweets of the day
1/
I'm sorry but this has gotta be a cash cow. What starts in Dubai always finds its way to America.
It's a robot manicure machine. This one is at an airport.
The nail salon in the same airport is $55. This machine? $17.
Gimme 14 of them right now.
— Chris Koerner (@mhp_guy)
5:09 PM • May 2, 2025
Insane arbitrage opportunity.
Women love to do nails and its astoundingly hard thing to do this properly. Trust me, I've tried to do for my mom & totally suck at it.
But this robot did it smoothly with appropriate proportions.
Solving this single problem alone is a $100m+ opportunity.
2/
This is your periodic reminder that Gaming is larger than the Movie and Music industries COMBINED
— Christian Keil (@pronounced_kyle)
5:03 AM • May 2, 2025
My thesis is Gaming makes you feel like a hero whereas when you watch a Movie, you are watching someone else play a hero barring some exceptional films like Rocky, Rudy, etc...
And gaming has mastered psychology. They have to. Otherwise the game will flop. You need the game to be simple but not too simple. Hard but not too hard. So the games that don't flop are the ones that get their players have high time spent.
And if you make a product that gets people to spend more time on it, then you can make more money from it. After all, we live in an attention economy.
3/
YAML > JSON when working with LLMs
Don't fry the model's brain with a curly-braces
— Ahmad (@TheAhmadOsman)
10:31 AM • May 3, 2025
I thought XML would be the best format for LLM prompting, but YAML is actually much better!
The reason is XML uses up way more tokens than YAML does. This matters unless you've got money to burn on tokens.
Ahmad pointed out a problem with JSON:
"generation entropy stability worsens over time while generating json. the model should not have to focus on the json curly-braces tokens order in its output instead of the real task on hand."
This ^^ seems like when we try multi-tasking and absolutely suck at both tasks.
Grok also mentioned this:
"For YAML vs. JSON, YAML's flexible syntax may keep entropy lower and more stable, using 48% fewer tokens, saving costs ($11,400/month for 1M GPT-4 requests)."
Just so you know - this mainly applies to complex prompts. For simple stuff, structured JSON might work better if all LLMs support it. But for smaller models, familiar formats like YAML might beat XML.
One more thing - JSON files can get huge! I've made simple test JSON files that grew beyond 10MB, but when converted to other formats, they shrunk in size dramatically.
Rabbit Holes
How To Take Control Of Your Own Destiny - George Mack by Chris Williamson
The Subtle Tricks Shopping Sites Use to Make You Spend More by Louise Matsakis
Navalmanack Bonus Section on BUILDING STARTUPS by Eric Jorgenson
What’d ya think of today’s newsletter? Hit ‘reply’ and let me know.
Do me a favor and share it in your company's Slack #marketing channel.
First time? Subscribe.
Follow me on X.
Reply